1 Definition and Intuition

A centroid is a point that characterizes the “average location” of a geometric object. Under an assumption of uniform density, it coincides with the center of mass of the object. More generally, centroids can be defined for distributions that assign different weights to different parts of a set—such as masses, probabilities, or importance values.

1.1 Center of mass connection

For a system of particles, the center of mass is the weighted average of particle positions. For a continuous body (like a lamina, wire, or solid), the same idea extends by replacing sums with integrals over the region. The centroid is the common geometric term used when the “mass density” is taken to be uniform (or when the weights represent density).

1.2 Balance-point interpretation

The balance-point picture is often used in statics. If a uniform planar lamina is supported on a frictionless pivot at the centroid, the torques caused by the lamina about the pivot cancel, yielding equilibrium. In symmetric situations, the centroid also matches the intuitive center suggested by shape geometry.

1.3 Centroid vs. other “centers” (e.g., centroid vs. centroidal axes)

“Center” is a broad label in mechanics and geometry. A centroid locates an averaging point of a given figure or distribution. By contrast, centroidal axes are coordinate axes chosen to pass through the centroid and simplify second-moment calculations, such as moments of inertia. Even more specialized notions include the center of figure (purely geometric center, when defined) and the center of inertia (linked to mass distribution and rotational behavior).

2 Centroid in Coordinate Geometry

In coordinate geometry, centroid formulas express the centroid coordinates in terms of given points, segments, areas, or assemblies of these primitives.

2.1 Centroid of a set of points

For an unweighted set of points \((x_i, y_i, \dots)\), the centroid is the arithmetic mean of the coordinates.

For points in the plane, the centroid \((\bar{x}, \bar{y})\) is \[ \bar{x}=\frac{1}{n}\sum_{i=1}^{n} x_i,\quad \bar{y}=\frac{1}{n}\sum_{i=1}^{n} y_i. \] This same structure holds in higher dimensions by averaging each coordinate component.

2.1.1 Weighted centroid for masses or importance values

When each point carries a weight \(w_i\) (such as mass or relative importance), the centroid becomes a weighted average: \[ \bar{x}=\frac{\sum_{i=1}^{n} w_i x_i}{\sum_{i=1}^{n} w_i},\quad \bar{y}=\frac{\sum_{i=1}^{n} w_i y_i}{\sum_{i=1}^{n} w_i}. \] A key feature is that negative or zero weights are generally not physical for mass, but they may appear in abstract weighting schemes (e.g., certain inclusion–exclusion constructions for holes).

2.2 Centroid of a line segment

For a uniform straight line segment, the centroid lies at its midpoint. If the endpoints are \(A(x_1,y_1)\) and \(B(x_2,y_2)\), then \[ \bar{x}=\frac{x_1+x_2}{2},\quad \bar{y}=\frac{y_1+y_2}{2}. \] This result reflects that every equal length element contributes equally to the “average” position along the segment.

2.3 Centroid of a polygonal area

For a planar region that is a polygon (or can be decomposed into polygons), the centroid can be computed by summing contributions of simpler subregions or using closed-form expressions based on vertex coordinates. One common approach decomposes the region into triangles with a shared reference point and aggregates their centroids weighted by their areas.

The centroid coordinates of a region \(R\) may also be expressed in terms of area moments: \[ \bar{x}=\frac{1}{A}\int_R x\,dA,\quad \bar{y}=\frac{1}{A}\int_R y\,dA, \] where \(A\) is the area of \(R\).

2.4 Centroid of a composite shape

For shapes formed by combining non-overlapping parts, the centroid can be obtained by treating each part as a “lumped” region. If subregions \(R_k\) have areas \(A_k\) and centroids \((\bar{x}_k,\bar{y}_k)\), then for the union (assuming the density is uniform within each part), \[ \bar{x}=\frac{\sum_k A_k \bar{x}_k}{\sum_k A_k},\quad \bar{y}=\frac{\sum_k A_k \bar{y}_k}{\sum_k A_k}. \] If the composite includes a hole, the hole is handled as a region with negative area contribution, which effectively subtracts its influence on the overall average.

3 Calculus-Based Formulation

When regions or curves have smoothly varying geometry, calculus provides compact definitions of the centroid as an integral “average.”

3.1 Centroid of a region using integrals

Let \(R\) be a planar region with uniform density. The centroid coordinates are \[ \bar{x}=\frac{1}{A}\int_R x\,dA,\quad \bar{y}=\frac{1}{A}\int_R y\,dA, \] where \(A=\int_R dA\).

If density varies, one replaces area \(dA\) with mass element \(\rho(x,y)\,dA\), yielding \[ \bar{x}=\frac{\int_R x\,\rho\,dA}{\int_R \rho\,dA},\quad \bar{y}=\frac{\int_R y\,\rho\,dA}{\int_R \rho\,dA}. \]

3.1.1 Derivation for planar lamina

The continuous analog of the mass-weighted average uses the idea that small area elements \(dA\) contribute proportionally to their mass \(\rho\,dA\). The total mass is \(\int_R \rho\,dA\), and the “moment” about the origin along the \(x\)-direction is \(\int_R x\,\rho\,dA\). Dividing the moment by total mass produces the centroid coordinate \(\bar{x}\). A parallel argument yields \(\bar{y}\).

3.1.2 Practical computation with area moments

In practice, many centroid calculations reuse tabulated formulas and precomputed area moments. Since \(\int_R x\,dA\) and \(\int_R y\,dA\) are moment-like quantities, centroid computations are often phrased using moment definitions relative to axes, or using transformed coordinates to reduce algebra. When the shape boundaries are simple, the integrals can be evaluated directly; otherwise, decomposition into primitives or numerical integration is used.

3.2 Centroid of a volume using triple integrals

For a solid volume \(V\) with uniform density, the centroid coordinates are \[ \bar{x}=\frac{1}{\mathcal{V}}\iiint_V x\,dV,\quad \bar{y}=\frac{1}{\mathcal{V}}\iiint_V y\,dV,\quad \bar{z}=\frac{1}{\mathcal{V}}\iiint_V z\,dV, \] where \(\mathcal{V}=\iiint_V dV\). Variable density \(\rho\) is handled by substituting \(dV\) with \(\rho\,dV\) in both numerator and denominator.

3.2.1 Symmetry shortcuts and simplifications

Symmetry can drastically simplify integrals. If a region is symmetric about the plane \(x=0\), then contributions from \(\pm x\) cancel, forcing \(\bar{x}=0\). Similar reasoning applies for \(\bar{y}\) and \(\bar{z}\). Rotational symmetry about an axis can imply equal centroid coordinates in perpendicular directions or place the centroid on the symmetry axis.

4 Parametric Curves and Centroids

Centroids extend beyond areas and volumes to wires and trajectories, where averaging is performed along a curve.

4.1 Centroid of a curve (wire) in 2D

For a planar curve parameterized by \(\mathbf{r}(t)=(x(t),y(t))\), the centroid of a thin wire with uniform linear density is found by averaging positions with respect to arc length. Using \(ds=\|\mathbf{r}'(t)\|dt\),

\[

\bar{x}=\frac{1}{L}\int x(t)\,ds=\frac{1}{L}\int x(t)\,\|\mathbf{r}'(t)\|\,dt,

\] \[

\bar{y}=\frac{1}{L}\int y(t)\,ds=\frac{1}{L}\int y(t)\,\|\mathbf{r}'(t)\|\,dt,

\] where \(L=\int ds\) is the total curve length.

4.2 Centroid of a curve in 3D

For a space curve \(\mathbf{r}(t)=(x(t),y(t),z(t))\), the same method yields \[ \bar{x}=\frac{1}{L}\int x\,ds,\quad \bar{y}=\frac{1}{L}\int y\,ds,\quad \bar{z}=\frac{1}{L}\int z\,ds, \]

with \(ds=\|\mathbf{r}'(t)\|dt\) and \(L=\int ds\).

4.3 Arc-length parameterization and weighting

Arc-length parameterization makes weighting uniform with respect to physical length, which is why \(ds\) appears in the formulas. If a curve is parameterized in a non-uniform way (e.g., by time \(t\)), the speed factor \(\|\mathbf{r}'(t)\|\) compensates so that equal physical lengths contribute equally. When a curve has non-uniform linear density \(\lambda(s)\), the integral weights become \(\lambda\,ds\), yielding mass-weighted averages along the curve.

5 Discrete vs. Continuous Centroids

Centroids can be computed for discrete sets (points or segments) or for continuous shapes (areas, volumes, curves). The relationship between these viewpoints is conceptually important for both theory and computation.

5.1 Relationship between summation and integration

Discrete centroid formulas are weighted sums: \[ \bar{x}=\frac{\sum_i w_i x_i}{\sum_i w_i}. \] Continuous formulas are weighted integrals: \[ \bar{x}=\frac{\int x\,\rho\,dA}{\int \rho\,dA} \] (or analogous expressions for volumes and curves). One can view integration as a limiting case of summation where the partition becomes finer and each weight corresponds to the mass of a small element.

5.2 Numerical approximation methods

When analytic expressions are hard or unavailable, numerical methods approximate the integrals or sums.

5.2.1 Sampling strategies and accuracy considerations

For regions, sampling can be done by:

  • Grid-based quadrature: evaluating \(x\) over a discretized mesh and summing contributions from cells that lie inside the region.
  • Monte Carlo integration: sampling random points in a bounding box and averaging contributions from points that fall within the region.
Accuracy depends on resolution, integrand variability, and how efficiently the method captures boundary geometry. For curves, numerical integration over parameter \(t\) requires careful inclusion of \(\|\mathbf{r}'(t)\|\) (or direct sampling in arc length) to avoid bias.

6 Common Applications

Centroids serve as compact summaries of geometry, enabling calculations that would otherwise require detailed shape descriptions.

6.1 Engineering mechanics and statics

In statics, centroids locate the effective application point of distributed loads under certain assumptions. For planar and three-dimensional bodies, centroid coordinates help determine the center of mass, which is central to balancing, stability analysis, and torque calculations.

6.2 Moments of area and second-moment context

While centroid itself is a location, it often appears alongside second moments of area (or mass moments). Moments of inertia about centroidal axes are frequently used in beam bending and vibration analysis because choosing axes through the centroid simplifies expressions and reduces coupling terms.

6.3 Computer graphics and shape analysis

In computer graphics, centroids support tasks like computing representative points for meshes, collision proxies, and layout of 2D shapes. For polygon meshes, averaging vertex positions may approximate a centroid, while more accurate approaches weight by triangle areas to better reflect the underlying surface distribution.

6.4 Data science: geometric averaging and barycenters

In data analysis, barycenters generalize the averaging idea for weighted sets. Centroid-like quantities appear in clustering interpretations (as means), in geometric feature aggregation, and in algorithms that combine shapes or embeddings by averaging coordinate systems or probability distributions.

7 Special Cases and Symmetry

Many figures have centroid locations that follow directly from symmetry, and some shapes require careful handling due to holes or composite boundaries.

7.1 Centroids of regularly shaped figures

For regular polygons, circles, rectangles, and ellipses, centroid positions often coincide with the geometric center. For example, a uniform rectangle’s centroid lies at the intersection of its diagonals, and a uniform disk’s centroid is at its center. For regular shapes, these results follow from rotational and reflection symmetries.

7.2 Using symmetry to reduce computation

Symmetry arguments reduce both dimensionality and integration complexity:

  • Reflection symmetry across an axis implies the centroid lies on that axis.
  • Rotational symmetry implies the centroid lies on the rotation axis.
  • Combined symmetries can fix multiple coordinates without integration.

Such shortcuts are particularly valuable for volumes and solids where direct triple integrals can be cumbersome.

7.3 Centroids of shapes with holes or cutouts

For a composite region with holes, the centroid is computed using the same area (or volume) averaging scheme, treating removed parts as negative contributions. If the removed region has centroid \((\bar{x}_h,\bar{y}_h)\) and area \(A_h\), then its contribution to the composite centroid is \(-A_h\bar{x}_h\) in the numerator and \(-A_h\) in the denominator. This approach extends to multiple cutouts by summing over all signed subregions.

8 Units, Conventions, and Coordinate Choices

Centroid computations depend on coordinate systems and units, but the meaningful geometric interpretation is coordinate-consistent.

8.1 Dimensional consistency

Centroid coordinates inherit the units of the underlying spatial dimensions. For instance, if \(x\) and \(y\) are measured in meters, then \(\bar{x}\) and \(\bar{y}\) are also in meters. Weighted averages do not alter dimensionality because weights represent either relative importance (dimensionless) or mass density elements that cancel between numerator and denominator.

8.2 Origin dependence and invariance properties

A centroid location depends on the chosen origin for its coordinate representation; however, the geometric point itself transforms predictably under coordinate changes. Under translations, centroid coordinates shift by the same amount as every point in the object. Under rotations, centroid coordinates rotate accordingly. In this sense, the centroid is invariant as a geometric object, even though its coordinate values change with the coordinate system.

8.3 Sign conventions and interpretation in different quadrants

Centroid formulas often involve coordinate components that can be positive or negative depending on quadrants. Negative centroid coordinates indicate the centroid lies on the corresponding negative side of an axis; they do not signal any problem with computation. Care is needed when interpreting sign with respect to the adopted axes, particularly in composite shapes where subtraction (holes) changes the effective contribution direction.

9 Worked Examples (Reference Calculations)

This section provides representative computations that illustrate standard centroid techniques.

9.1 Centroid of a triangle

For a triangle with vertices \(A(x_1,y_1)\), \(B(x_2,y_2)\), and \(C(x_3,y_3)\), the centroid is the average of the vertex coordinates: \[ \bar{x}=\frac{x_1+x_2+x_3}{3},\quad \bar{y}=\frac{y_1+y_2+y_3}{3}. \] This holds for a uniform triangular lamina.

9.2 Centroid of a rectangle with a removed subregion

Consider a rectangle of area \(A_1\) with centroid \((\bar{x}_1,\bar{y}_1)\), from which a rectangular hole of area \(A_2\) with centroid \((\bar{x}_2,\bar{y}_2)\) is removed. The composite centroid is \[ \bar{x}=\frac{A_1\bar{x}_1-A_2\bar{x}_2}{A_1-A_2},\quad \bar{y}=\frac{A_1\bar{y}_1-A_2\bar{y}_2}{A_1-A_2}. \] This example demonstrates the negative-area convention for holes.

9.3 Centroid of a composite line or area

For a composite line made of two segments with lengths \(L_1,L_2\) and centroids (midpoints) at \(\bar{x}_1,\bar{x}_2\) along a chosen axis, the centroid coordinate is \[ \bar{x}=\frac{L_1\bar{x}_1+L_2\bar{x}_2}{L_1+L_2}. \] For a composite area, the same pattern holds with area \(A_k\) replacing length, and each subregion’s centroid \((\bar{x}_k,\bar{y}_k)\) contributing in proportion to its size (with negative sign if the subregion is removed).

10 Limitations and Edge Cases

Certain situations require care to avoid misapplication of centroid formulas.

10.1 Non-uniform density assumptions

All centroid formulas presented with simple geometric averaging presume uniform density (or, in the weighted case, correctly specified weights). If density varies and the variation is not captured by the weighting, the computed point will not represent the center of mass.

10.2 Degenerate shapes (zero area/length)

Centroid definitions for continuous objects require a nonzero total measure: area for planar regions, volume for solids, and length for curves. If a shape has zero area (e.g., a line with no thickness treated as a region) or zero length (a point-like object), the centroid is undefined under the standard averaging framework.

10.3 Handling coordinate singularities and numerical stability

Numerical approaches can suffer from stability issues when regions are described implicitly (e.g., via equations that create near-singular behavior) or when sampling is too coarse near boundaries. For parametric curves, evaluating \(\|\mathbf{r}'(t)\|\) may be sensitive to parameter values where the derivative is small or noisy. Robust implementation often uses adaptive sampling, careful parameter scaling, or subdivision strategies.

Centroids connect to other “average” or “center” notions that arise in geometry and mechanics.

11.1 Barycenter and mass moments

Barycenter is a general term for weighted averages of points, closely aligned with centroid in many contexts. Mass moments provide complementary information: they quantify how a distribution spreads about axes and are used to compute inertial effects and centroidal second moments.

11.2 Center of figure vs. center of mass

The center of figure is a purely geometric notion, often defined for symmetric shapes. The center of mass accounts for mass distribution and may differ from the center of figure when density is non-uniform or when the shape is not uniform.

11.3 Center of inertia and rigid-body generalizations

In rigid-body dynamics, the “center of inertia” is sometimes used in relation to how mass influences rotational behavior, such as through moments and products of inertia. While centroid gives a translational averaging point, inertia-related centers and tensors describe resistance to rotation, which can depend on the orientation and distribution of mass.